Going Beneath the Surface

In medicine, highly targeted treatments have great potential to improve outcomes, while simultaneously reducing recovery times and the risk of potentially serious complications. Whether the goal is to collect a biopsy sample from tissue deep within the body or deliver chemotherapy directly to a tumor, minimally-invasive targeting of the tissue of interest is the best option available. With the advent of cutting-edge technologies such as robotic surgery and advanced imaging techniques, the precision of medical interventions has reached unprecedented heights. Surgeons can now navigate intricate anatomical structures with enhanced accuracy, ensuring minimal damage to surrounding healthy tissues and organs.

But to date, these procedures have always involved manual teleoperation. Despite the many successes associated with these procedures, their limitations have become apparent. There are many anatomical obstacles deep within the tissues of the body that can present a surgeon with a high level of uncertainty about the best course of action. Additionally, due to natural motions of the tissue as the patient breathes, or their heart beats, for example, navigation of tiny internal structures can be exceedingly challenging. Moreover, the cameras, actuators, and other hardware required to support remote teleoperation increases the size of the device, which in turn makes it difficult or impossible to safely reach many regions within the body.

The diminutive size of needles makes them ideal for many interventional medical procedures, however, safe and accurate navigation of needles through the body is not always possible. But these present difficulties may soon be a thing of the past, thanks to the work of a team led by researchers at the University of North Carolina at Chapel Hill. They have developed a needle-like medical robot that can, unlike any other such robot, operate completely autonomously. By harnessing the power of AI algorithms for navigation, the robot can navigate with incredible precision to safely access areas of the body that were previously too challenging to reach.

The researchers developed a laser-patterned, flexible needle-robot that can move along curvilinear trajectories. Traditional imaging techniques, like CT scans, provide the system with a highly detailed three-dimensional map of the patient’s body. Leveraging this information, a machine learning algorithm calculates the best path available to autonomously drive the robot to its destination. The algorithm takes into account any anatomical obstacles or other important structures like significant blood vessels and will avoid traveling too close to them. Natural body motions, like those caused by breathing, are even factored into the equation, such that the robot will pause and wait for a safe window of insertion before progressing through a sensitive region.

To test the performance of their device in a real-world setting, the robot was evaluated in pigs. The needles were evaluated in one of the most challenging regions of their bodies — the lungs. Due to the constant motion of lungs during breathing, this test would serve to highlight the benefits of this new robot over existing techniques. In a number of trials, it was found that the robot could safely reach its target with errors in accuracy that were less than the radius of clinically relevant lung nodules.

The leap from experiments in pigs to performing procedures in humans is by no means small and will not happen overnight. But the team is diligently working towards that goal, continuing to refine their device so that it might be able to handle the most challenging of cases one day, while also offering unprecedented new guarantees of patient safety. Looking even further down the road, the researchers hope to leverage their technology to build more types of autonomous robots that can assist surgeons in their work.